Electrical safety around the pond

[seanmckinney]

Perhaps this could be moved to the library after it has run its course as an active thread. If the thread is deemed worthwhile to go into the library I would prefer to up load the images onto koiphen rather than leave them on my freeserve page. I know I can up load them via 'attachment' but I have the impression I then loose control of where in the post the images appear. Is that correct and if so who do I upload them to specific points in the post.

I am told that the message of this post is not clear, so here it is before you get lost in detail.



A GFI will only trip if there is a function ground circuit connected to the device or pond etc.


Recently there have been what I would regard as a few misleading posts about electrical safety around the pond. Before these I had said I would do an 'experiment' to demonstrate some points, well I finally got off my rear end and did the experiments.

DO NOT ATTEMPT TO RECREATE THE EXPERIMENTS IN THESE PHOTOS UNLESS YOU KNOW WHAT YOU ARE DOING.

It would actually be useful if someone Stateside who DOES KNOW WHAT THEY ARE DOING did recreate them with a US spec GFI, I think your 'lightest' trip is 5mA, whereas ours, as far as I know, is 30mA and the one used in this experiment is a '30mA' trip. I actually checked it a while a go and from memory it trips somewhere around 25mA, which is good.

You should note I am wearing two layers of latex gloves and on my feet all through the experiment wearing rubber wellies.

Ok
Photo 1) the apparatus, insepection lamp on, a plastic bucket, to simulate the pond liner, full of pond water
an inspection lamp to show that the power is there
suitable extension cables

The power comes from the cable with the black socket and the opened plug, the extension that starts with the open plug ends in the white socket and the socket feeds the inspection lamp.




Photo 2) the earth/ground wire in the open sockey has been disconnected, sorry it is blurred I didnt realise the camera had focused on the meter rather than the plug. This mean that the inspection lamp is unearthed/ungrounded.




Photo 3) the white plug and socket in the bucket of pond water with the light still lit, note no RCD/GFI protection at the moment on the inspection lamp





Photo 4) plug and socket in the water, lamp lit and meter showing the voltage between the water 4 or 5 inches from the live/hot socket and electrical earth/ground. 121.5V. Note the supply voltage was around 238V during this experiment.







Photo 5) RCD/GFI added BUT the earth/ground to the lamp is STILL disconnected, the 2nd meter is showing the voltage between the water and the acual wet soil under the bucket 118V, I cant read the other meter, sorry

It doesnt show very well but the lamp is still fully lit, IE, THE RCD/GFI DID NOT TRIP!!!!!!!!!!!!!!!!!!!!!

The above is the situation present with a faulty two wire device with no grounding of the water EVEN IF A GFI IS PRESENT


Photo 6) a clear shot of the voltages between the water and electrical earth/ground and between the water and the soil





As soon as I touched the disconnected earth/ground wire to is pin the RCD/GFI tripped, ie the electrical earth/ground from the lamp carried some of the current otherwise destined for the neutral and thereby creating a 25mA+ difference between the current flowing in the live/hot and neutral circuits within the RCD/GFI and tripping the RCD/GFI

[seanmckinney]

Photo 7) Earth/ground again disconnected but the water earthed/grounded through the meter, which is set to measure current, to the soil. The meter set that way is effectively a direct short, the current shown is 16.8mA, the RCD/GFI did not trip.

It would have with a 5mA tripping GFI.




Photo 8) RCD/GFI removed but the earth/ground reconnected, the plug and socket are still in the water and no fuse blew

The eagle eyed may spot that the cable is old, old colour scheme, but it is still good.

[cbcdesign]

I assume from your post Sean you are saying that a more sensitive RCD is required for pond use. If so, Its not a bad idea if you tend to put your hands in the water with the mains supplies connected like I do. That said, I pulled a liver submersible out of my pond a couple of summers ago, my 30mA rcd suddenly tripped and I didn't feel a thing so even a lower sensitivity RCD is worth having.

You probably know that your Avo has an internal shunt for current measurements. The shunt resistance would limit fault current of course, hence the fact that your RCD didn't trip with the meter in circuit.

[chipmaker]

Sean
Do you recall a similar post or experiment that I did on the GW forum in the thread concerning 2 prong versus three prong devices with a fellow named Horton? I have a different user name on GW as old Spike booted me for mention of another website on more tha one occasion......My findings were similar to yours. I would caution about using regular latex gloves etc, as all rubber gloves and boots etc are not created equal....... All it takes is a miniscule path for moisture/ current flow and it could ruin your day. Just be carefull with these experiments.

You mentioin old color codes.......I think I have the color comibinations down now, as 9 ouit of 10 items yuou buy here is made in another country and they use different colors than the red/white/black/green the USA does. Is the blue=hot, brown=neutral, and green w/wo/yellow stripe = ground on your color codes? What would the color be for the conductor if its 220 VAC? Red perhaps?

[cbcdesign]

Brown is Live, Blue for Neutral, Green/yellow is Earth.

[seanmckinney]

CDC those posts are primaily meant to report and show 'experimental' observations, I have the impression that sometimes 'myth and legend' are given more credability than fact but, if it is accepted that the photos havent been faked then, it is pretty hard to argue against the results demonstrated in the above posts.
I think those posts clearly show that in order for an RCD/GFI to protect a pond, then that pond or the appliances within or connected to that pond must have GOOD earths/grounds. Ungrounded/unearthed appliances with no path or a poor path to ground/earth for any 'errant' electricity will not be protected by a GFI/RCD or by a fuse/circuitbreaker.
With regard to "more sensitive RCD is required for pond use", I suspect that is the case but that portion of those posts is more for the US 'side' of 'things' since ungrounded devices are available there, I have yet to knowingly see an unearthed appliance for pond use here in the UK. As such 'remote grounding' is more common in the US and since water would seem to be resistive it would seem sensible to reduce the leakage current required to trip an RCD. Whether or not a more sensitive GFI, eg 5mA trip, would protect a pond if the voltage source and grounding point were seperated by some distance, ie several feet, I do not know and that is my concern.
What I couldnt show was that to get a leakage current even close to the trip value for my RCD I had to have the point of the probe almost in the socket's live/hot 'hole', I didnt want to put the probe actually into the live hole incase the probe would touch the live/hot contact which would have damaged the meter.
I did wonder about the resistance of the shunt on the below-10A-ranges and will repeat that experiment with a wire short, however I did initially test it on the 10A range/connections and the shunt on that, from memory, has a very low value.

Chipmaker, sorry I dont remember that post but I do remember some debates on AWGS and GW that got heated and I have seen one or two things on Koiphen that are IMO questionable. I am posting this on Koiphen because koiphen has the longer lasting archive, Awgs lost that facility when they upgraded and that I am aware of GW has never had a long lasting archive.
There is only 1 voltage available here for 'domestic use' and that is now 230V at 50Hz, it used to be 240V at 50Hz. I believe the voltage was lowered to match that used in Europe. The current colour scheme is brown=live/hot, blue=neutral and green&yellow=earth/ground, I think the old colour scheme was red=live/hot, black=neutral and green=earth/ground, I think it was changed to make things safer for those people with colour blindness.

I hope that this thread will not turn into a debate as, to me, with the exception of the remote grounding and more sensitive GFI nothing in the results would appear debatable.

[MCrossley]

Thank you for posting this; as this is what I was trying to get across to people on the "GFI Tripped" thread. It realy makes me wonder how so many people survive or their property for that matter in how uneducated they are when doing a electrical project. As said earlier this stuff can and in most cases will kill you in short order and if you are not comfortable or at all question your ability to safely do an electrical project DON'T!!! I don't say this to offend anyone but you guys and gals on this board are all great people that would be nice to keep around for a while. The point that he is trying to make in this thread is that like I said in the other thread a GFI is meant to measure an imballance of current between circuits and also ground if ground is taken away it has no reference for the GFI to trip thus causing electrical shock. Grounding devices is the best and only way to safely wire things in any application especially outsoors or near water.


Anyhow enough preaching but wanted to say thank you for others caring about others....

[Kent Wallace]

I regularly see many ponds that have the electrical as an afterthought. My "Charles Pond retro" thread was an electrical nightmare and I should have taken before and after photos of the wiring mess left by the previous landscaper.
I advocate hard wiring the plug ends of pump chords into a sealed box with outdoor switches and no exposed plugs near the pond.

Mcrossley, On your way through to (somewhere in NW AZ?) please come
visit us in Vegas.
This is a pic of as close as I like to get given the space provided.

[MCrossley]

sounds great I appreciate the offer not sure when I am going to drive through yetl I have come down with some serious medical problems that has me a bit on a short leash thats why I have been able to post regularly otherwise I would be at work :-)

[Johns]

Sean suggested that someone in the USA duplicate his experiment using a 5Ma trip current GFCI. I have attempted to do so.



First, since pure water is not a conductor, I used water from one of my fish ponds to more closely approximate the electrolyte properties found in most home water gardens. I do not salt my ponds, but be advised that if your pond water IS salted, it will be even a better conductor of electricity that that used in the experiment.



Second, I would like to point out that proper installation wiring of a GFCI is important, as if wired improperly, it will provide no protection. If you are installing your own GFCI’s, be sure to test them by use of the test button before using. If you press the test button and the GFCI does not trip, you either have a bad GFCI or you wired it wrong.



In my rendition of Sean’s experiment, I created two test setups: one without a GFCI and one with a GFCI. With power off, I fed a three wire cable, Hot wire (Black) neutral (White) and equipment ground (earth) to a receptacle immersed in a bucket of water and a plug in the receptacle leading to a trouble light. I included two switches, one in the equipment ground conductor before the immersed receptacle and the other in series with a separate conductor run from the bottom of the bucket to a ground stake embedded in the earth. Setup number two includes a Leviton 8599-GY 5 Ma trip GFCI (See http://www.levitonhelpdesk.com/catalog/ and enter “8599-GY” in the catalog number search box and click “go”) before the equipment ground switch. A work light was plugged into the receptacle and the receptacle was immersed in the water in the bucket. A schematic of the arrangement:







Here are the protocol and test results:





CONCLUSIONS:

Either the presence of an equipment ground OR a grounded pool will cause a tripped GFCI when there is exposure of the line voltage to the water. Of course for the former to be true, the equipment ground must also be exposed to the water. The design of the device may or may not accomplish this end.


Unfortunately, I do not possess test equipment that will measure low AC amperage (my meter only goes down to 5 amp range). I did measure the DC resistance between the immersed receptacle and the earth ground conductor through the water and in the test configuration the resistance was only 450 Ohms. At 120 volt potential, this would allow amperage of .267 AMPS and at the measured 60 volt range, .133 AMPS, either of which could kill a human. Of course the actual resistance is dependent on the electrolyte properties of the water and the distance from the fault to the pool grounding conductor.



I would suggest that Sean’s experiment is easily replicated and is therefore scientifically sound.

Johns

P.S. Much thanks to Pandora in educating me to the edit function.

[seanmckinney]

Thanks Johns.

I havent had a chance yet to repeat the experiment with a direct short to the soil, it was my mum's 90th today and I have been running family left right and centre, or cooking etc, the results of which are probably riskier than the above experiment, since Friday and there has been raining and I aint trying this in the rain.

[Keiko, The Killer Koi]

Great Thread!